EP1206044B1

EP1206044B1 - Matched filter and correlation detection method

Info

Publication number: EP1206044B1
Application number: EP20010126302
Authority: EP
Inventors: Takenori Matsumoto; Naoshige Kido
Original assignee: Matsushita Electric Industrial Co Ltd
Current assignee: Panasonic Holdings Corp
Priority date: 2000-11-09
Filing date: 2001-11-06
Publication date: 2006-09-06
Anticipated expiration: 2021-11-06
Also published as: JP3415579B2; US20020054625A1; EP1206044A3; EP1206044A2; JP2002152084A; DE60122848T2; DE60122848D1; US6985517B2

Description

BACKGROUND OF THE INVENTION

Field of the Invention

The present invention relates to synchronization acquisition processing according to a spectrum direct sequence communication system, and more particularly, to a matched filter and correlation detection method used in preamble searching.

Description of the Related Art

A spread spectrum communication system (CDMA communication system) features intrusion resistance, high level of signal concealment and high frequency utilization efficiency, etc. and is used for mobile communications and space communications, etc.
When a mobile station (mobile terminal) requests a base station to start a communication, a cellular phone system based on a CDMA communication system transmits a preamble signal.
The preamble signal is a signal obtained by, multiplying a known signature code (original signal) with periodicity by a common spreading code. ,
The signature code (original signal) is a code with a total length of 4096 symbols made up of a 16-bit basic data bit pattern appearing 256 times consecutively.
Here, 16 types of basic data bit pattern are provided. Therefore, as a result, there are also 16 types of signature code. Each signature code is used to identify a reception group and the mobile station side determines which signature code should be used as appropriate.
The mobile station that requests transmission to the base station selects any one of the above-described 16 types of signature code, spreads/modulates and sends this signature code to the base station as a preamble signal.
When a signature code (original signal) is spread/modulated, the signature code is multiplied by a common scrambling code (common spreading code).
The base station receives a spread spectrum signal in which the preamble signal is inserted, despreads it and detects the signature code (original signal). Thus, the base station recognizes the communication request from the mobile station and determines the mobile station with which communication is allowed to start.
The information on the signature code detection timing is used as the timing information to establish synchronization for communication with the mobile station with which communication is allowed to start.
The base station needs to instantaneously detect communication requests from many mobile stations that exist in a cell and determine to start a communication.
A digital matched filter (hereinafter simply referred,to as "matched filter") is used to detect a signature code.
As a metho,d for carrying out high-speed correlation detection at the base station, a method of detecting correlation is available whereby 16 matched filters are placed in parallel, multiplications are performed on 16 types of preamble signal in parallel (that is, simultaneously), the results are integrated in the time direction and a correlation is detected for every preamble signal (that is, for every signature code).
This configuration is shown in FIG. 9. FIG.9 shows a case where 16 correlation detectors A1 to A16 are provided and each detector detects a correlation with one of signature codes (1) to (16).
Despreading code generation sections 104a and 104b each output data strings (that is, preamble signals) obtained by multiplying signature code (1) or signature code (2) by a common spreading code as despreading codes.
However, in the case of a circuit (FIG.9) based on a system whereby 16 matched filters are placed in parallel and correlations of 16 types of preamble signal are detected simultaneously, the circuit scale increases significantly. This is accompanied by an increase of power consumption of the circuit.
Furthermore, the matched filter may be used not only for preamble searching but also for acquisition of synchronization and synchronization follow-up or RAKE combining and path determination processing for communications with directivity.
Thus, adopting a hardware configuration specialized only for preamble searching may involve a problem that that configuration will lack flexibility in the case of performing other processing and increase unnecessary parts of the circuit.
The present invention has been implemented focused on such problems and it is an object of the present invention to simplify the configuration of the matched filter section in particular to carry out high-speed preamble searching while keeping its flexibility and at the same, time reduce both the circuit scale and power consumption.
Document "Clarifications on Golay-Hadamard Sequence Based RACH Preamble" http://www.3gpp.org/ftp/tsg-ran/WG1/TSGR1_06/Docs/Pdfs/r1-99893.pdf, XP002174961 introduces the structure of the Golay-Hadamard preamble and the corresponding detector at the base station. For Golay-Hadamard preamble, the correlator coefficients on-line are generated. This can be done by running the address bus switching at chip rate Fc and the counter at clock rate of Fc/16. Eight level of adder trees and sign multipliers are needed for the major complexity part of the correlator.
Document "Proposal for RACH Preambles" http://www/3gpp.org/ftp/tsg-ran/WG1-RL1/TSGR1_07/Docs/Pdfs/r1-99C28.pdf, XP002260914 discloses that the preambles consist of length 16 signature sequences formed from a set of orthogonal Gold codes spread by a 4096 chip segment section. The document also describes several possible preamble detection schemes, namely―Coherent Accumulation, Non-coherent Accumulation and Differential Detection. The detection by Coherent Accumulation can be performed by summing the matched filter outputs and squaring the result to give the decision statistic. Alternatively, in a Non-coherent Accumulation, matched filter outputs can be accumulated within some number of segments of the preamble, the result squared and then accumulated. Differential Detection may be performed by accumulating within a segment and then taking the conjugate product of consecutive sums.

SUMMARY OF THE INVENTION

The present invention, focused on the periodicity of a signature code (original signal) which serves as the source for creation of a preamble signal, replaces the conventional method of carrying out simultaneous and parallel correlation detection processing by a method of processing on a time-division basis (time-division system) to reduce the scale of the circuit.
That is, when a reception signal is despread and a correlation is detected, a prior art multiplies the reception signal by a common spreading code and signature code and performs a cumulative addition on sets of symbols of the resulting signals and detects a correlation.
That is, the prior art carries out correlation detection using the procedure shown in expression (1) below. $≪ (common spreading code \times signature code) \times reception signal ≫$

where << >> means an operation of cumulative addition of a reconstructed symbol.
In contrast, the present invention first multiplies the reception signal by a common spreading code and then carries out a first cumulative addition. Then, the present invention multiplies the result of the first cumulative addition by a signature code and performs a cumulative addition on sets of symbols of the resulting signals (second cumulative addition).
That is, the correlation detection procedure of the present invention is as shown in expression (2) below. $≪ ≪ (common spreading code \times reception signal) ≫ \times signature code ≫$

where << >> means an operation of cumulative addition of a reconstructed symbol.
The calculation result according to the procedure in expression (1) above is equivalent to the calculation result according to expression (2).
Thus, the present invention does not perform calculations for correlation detection at a stroke but performs calculations step by step at a plurality of times on a time-division basis and carries out a first cumulative addition at some midpoint of processing focused on the periodicity of a signature code.
Using such a correlation detection method based on a time-division system requires the first despreading processing only to multiply the reception signal by a common spreading code, which allows the first despreading processing to be carried out by one (common) multiplication circuit and suppresses the circuit scale from increasing.
Furthermore, carrying out the first cumulative addition reduces the number of symbols, for example, to 1/256, alleviating the burden on the data processing thereafter.
Then, the result of the first cumulative addition is multiplied by a signature code and finally a second cumulative addition is performed to detect a correlation, which allows extremelyefficient and high-speed correlation detection processing using minimum hardware most effectively.
Adding some contrivance such as applying parallel processing using a plurality of multipliers to multiply the signature code further, improves the processing efficiency.
Improving the circuit configuration of multipliers themselves or using a minimum volume of memory on a time-division basis, etc. can further reduce the circuit scale or simplify the circuit design.
In a preferred mode of the present invention, a common spreading code (despreading code) is multiplied ion the reception signal first. From the bits that make up the resulting signal, a set of data bits corresponding, to the cycle of the signature codes (original signals) which are relatively in a same positional relationship are extracted. Then, a cumulative addition (first cumulative addition) is applied to every set of the extracted bits. Then, a signature code (original signal) is multiplied and cumulative additions are performed on symbols (second cumulative addition)
It is possible to reduce the circuit scale significantly by separately carrying out multiplication of a spreading code and multiplication of a signature code (original-signal), and carrying out cumulative additions in two stages using the periodicity of the original signal.
The effect of the ability of the present invention to carry out calculations with a smaller circuit scale than the conventional art becomes more outstanding as the number of original signals to be calculated (types of preamble signal) increases.
Furthermore, even when codes are multiplied, adopting a system of selecting an input of the selector according to the code pattern makes it possible to perform, programming freely so as to match the code pattern, thus securing the degree of freedom of the circuit.

BRIEF DESCRIPTION OF THE DRAWINGS

The above and other objects and features of the invention will appear more fully hereinafter from a consideration of the following description taken in connection with the accompanying drawing wherein one example is illustrated by way of example, in which;

FIG. 1 illustrates a configuration example of a matched filter of the present invention;
FIG. 2 is a flow chart to explain an operation of the matched filter shown in FIG. 1 ;
FIG.3 is a block diagram showing another configuration example of the, matched filter of the present invention;
FIG. 4 illustrates a signal waveform to explain a main operation of the circuit shown in FIG. 3;
FIG. 5 is a block diagram showing another configuration example of the matched filter of the present invention;
FIG.6 is a block diagram showing another configuration example of the matched filter of the present invention;
FIG. 7 is a block diagram showing a configuration of a CDMA communication base station apparatus of the present invention;
FIG. 8 is a flow chart showing a main operating procedure of the base station apparatus shown in FIG.7;
FIG.9 is a block diagram showing a configuration of a matched filter (conventional example) not using the present invention; and
FIG. 10 illustrates signal waveforms to explain a correlation detection method (conventional example) by in-phase addition calculation using signature codes (known periodic signals).

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

With, reference now to the attached drawings, embodiments of the present invention will, be explained in detail below.

(Embodiment 1)

The matched filter of the present invention is used when a CDMA communication base station apparatus, receives a signal sent from a mobile station and carries out a preamble search.
Before explaining a specific configuration of the matched filter, an overall configuration of the CDMA, communication base station apparatus will be explained using FIG. 7. As shown in FIG. 7. a QPSK (Quadri-Phase Shift Keying) signal, received by antenna AN of base station 400 is demodulated by demodulation circuit 401. As a result, an I (in-phase component) and Q (quadrature component) signals are output.
The I and Q signals are converted to digital signals by A/D convertor, 402 and given to matched filters 403a and 403b provided for respective signal.
Matched filter 403a is a correlation detector for processing of the I signal and matched filter 403b is a correlation detector for processing of the Q signal.
A configuration and calculation me,thod of the matched filters will be described in detail later. Matched filter 403a (also matched filter 403b) stores data in reception data storage section 500 temporarily.
A despreading code (same as spreading code and may also be described as a spreading code) generated from despreading., code generation section 501 is temporarily stored in despreading code storage section 502.
Then, at multiplication o,pe,ration, section 503, the despreading code is multiplied on the reception data and a cumulative addition (in-phase addition) is performed according to the cycle of a signature code (first in-phase addition) and the result is stored in memory 505.
The data read from memory 505 is multiplied by, for example, each of 16 types of code generated by signature code generator 506 and the data,of each symbol resulting from the multiplication is subjected to a cumulative addition (in-phase addition) (second cumulative addition). In this way, correlation detection is performed on the preamble signal included in the reception signal.
The I and Q correlation values (correlation signals) output from matched filters 403a and ,403b are input to power calculation section 601 via interpolation filter 600. Then, the reception power is calculated, as I²+Q².
Then, averaging section 602 averages the calculated reception power over a predetermined time width (window width taking into account a propagation delay required for a preamble signal to propagate from the mobile station to the base station).
Then, peak detection section 603 compares the averaged power with a threshold and detects peaks. In this way, a delay profile is created.
From the delay profile, it is possible to detect which signature code has been received with what amount of delay.
When it is detected, as a result of creation of the delay profile, that a preamble signal has been transmitted, the base station starts control (demodulation search) to establish communication synchronization with the mobile station in order to give communication permission.
Timing control for synchronization establishment is carried out by tracking section 604 and timing control section 605.
The code generation timing of spreading code generation section 606 is controlled by a timing control signal output from tracking section 604 and the reception signal is despread by despreading section 607 then subjected to coherent detection and RAKE combining by coherent, detection section 608 and RAKE combining section 609, respectively.
This is the; explanation about the overall configuration o,f, the, base station system. Then, correlation detection regarding a preamble signal which is inserted in the reception signal using the matched filter of the present invention will be explained.
First, a basic configuration of the matched filter of the present invention and a basic procedure for detecting a correlation regarding a preamble signal will be explained using FIG.1.
The mobile station 1 shown in FIG.1 selects one of 16 types of signature code (original signal) Sn (n :one of 1 to 16) for transmission.
The signature code is a periodic code in which a basic pattern made up of "1" and "-1" is repeated .
To facilitate the understanding, suppose, in FIG. 1, a spreading factor is "1" and the basic data bit pattern (hereinafter referred to as "basic pattern") is made up of 4 symbols (4 chips). That is, in FIG.1, the basic pattern consists of 4 symbols of "ab, bn, en and dn (n: one of 1 to 16)". Each of "ab", "bn", "cn" and "dn" takes a value of "1" " or "-1" .
In this way, the periodic signature code using such a basic pattern (4 symbols) as a unit is multiplied by a common spreading code (scrambling codes "K1, K2, K3, K4 ··· ···" ) and thereby a preamble signal is created.
In FIG. 1, each bit of the preamble signal is expressed by "A" to "L".
As described above, the spreading factor is "1" here. That is, a one-symbol period of the signature code is the same as a one-chip period of the, spreading code. Therefore, one symbol is equivalent to one chip. In, the following explanation, a term "symbol" will be used to refer to the data before being multiplied by a spreading code.
The preamble signal is multiplexed with a transmission signal and transmitted. Base station apparatus 2 carries out despreading by multiplying the preamble signals (A to L) included ip the reception signal by scrambling codes (K1 to K12···). If synchronization is established, the signature codes (Sn) are reconstructed as a result of despreading.
Then, a set of symbols placed at relatively equivalent positions corresponding to the cycle of signature codes are extracted and a first in-phase addition is carried out.
More specifically, for data "an" of chip number ① shown at the, top of FIG.1, data "an" of chip numbers ⑤ and ⑨ in a positional relationship corresponding to the cycle of signature codes are extracted and this set of "an"s are subjected to accumulative addition by cumulative, adder 10a.
Likewise, data "bn" of chip numbers, ②, ⑥ and ⑩ are subjected to a cumulative addition by cumulative adder 10b.
Likewise, data "cn" of chip numbers ③, ⑦ and ⑪ are subjected to a cumulative addition by cumulative adder 10c.
Data "dn" of chip numbers ④, ⑧ and ⑫ are subjected to a cumulative addition by cumulative adder 10d.
Thus, a: set of "an"s, a set of "bn"s, a set of "cn"s and a set of "dn" s are extrarcted and the first cumulative addition section 3 (adders 10a to 10d) executes cumulative additions (first in-phase addition).
When described, more exactly, adder 10a performs a cumulative addition on the same number of "an" s as the number of times the basic pattern is repeated, ("an" s positioned at the first symbol of the basic pattern). Here, if a correlation is established, adder 10a outputs 3an.
Likewise, adders 10b, 10c and 10d perform a cumulative addition (in-phase addition) of a set of "bn" s, a set of "cn" s and a set of "dn" s, respectively.
Then, multiplication section 4 (provided with multipliers 12a to 12d) multiplies each of 16 types of signature code Sn sequentially at different times.
Then, second cumulative adder 5 (provided with adder 13) performs a cumulative addition on symbols of a symbol string obtained as a result of multiplication (second cumulative addition) for every signature code. The second cumulative addition adds up 4 lines of data.
The circuit in FIG. 1 performs calculations for a total of 12 symbols. Therefore, if a correlation is established, a correlation value "+12 (= 3 x 4) is output. Correlation detection is performed in this way.
In this way, the matched filter in FIG.1 performs the following, characteristic operations:

(1) First, a common spreading code is multiplied and then a signature code (original signal) is multiplied. That is, codes are multiplied not at a stroke but on a time-division basis.
② After a common spreading code is multiplied, a set of data bits (for example, a set of "an"s) in a relatively, equivalent postiotional relationship focused on the periodicity of signature codes (original signals) are subjected to a cumulative addition (in-phase addition) to reduce the number of symbols. In FIG.1, the number of symbols is reduced,to "4" (equal to the number of symbols of the basic pattern).
③ Then, signature codes Sn are sequentially multiplied at different times to detect a correlation. Then, a final cumulative addition (in-phase addition) is performed for every signature code. That is a cumulative addition (in-phase addition), is also carried out on a time-division basis.

That is multiplying a code obtained by multiplying a spreading code by a signature code on the reception signal and then subjecting each symbol to a cumulative addition (conventional system) is equivalent to multiplying the reception signal by a spreading code and thereby performing a cumulative addition and then multiplying the signature code and finally performing a cumulative addition on each symbol (time-division system of the present invention), and the results obtained are the same.
Thus, an outstanding feature of the processing of the present invention consists in adopting a concept of carrying out multiplications in different steps in the time direction and a concept of carrying out a first cumulative addition at some midpoint of processing focused on the periodicity of signature codes.
Here, a correlation detection method using an in-phase addition (cumulative addition) will be explained briefly.
An in-phase addition (cumulative addition) is a calculation method used to create a delay profile (calculation, of reception power) using known symbols (preamble signals and pilot signals) inserted for initial synchronization and synchronization follow-up, when two lines of mutually orthogonal reception signals (I component, Q component) such as QPSK exist.
That is, in a calculation of reception power, instead of directly calcul,ating I² + Q² and carrying out averaging processing, this is a method of carrying out cumulative additions of (I+I+I···+I) and, (Q+Q+Q+···Q) about a plurality, of continuous in-phase symbols and adding up the squares of these cumulative, addition results.
Adopting such a method has advantages that the absolute values, of signals increases through cumulative additions, noise signals, randomly superimposed on the reception signals, are cancelled out through addition operations, thereby improving the accuracy of calculations of reception power.
A specific example of in-phase additions is shown in FIG.10.
The signal waveform shown at the top of FIG.10 shows a signature code with a cycle of 32 symbols.
Multiplying this signature code by a common scrambling code (spreading code) creates a preamble signal as described on the second row of FIG.10 and this becomes a transmission signal (and reception signal).
Multiplying the reception signal by a scrambling code (spreading code) on the receiving side reconstructs (reproduces) the signature code as shown in the signal waveform described on the third row of FIG.10.
Then, multiplying on the receiving side the same signature code as t,hat multiplied on the transmitting side at synchronized timing gives all symbol data placed on the positive side as shown by the signal waveform described on the fourth row of FIG.10.
Carrying out an in-phase addition (cumulative addition) in this state outputs a correlation peak value (+32) and detects a correlation as shown by the signal waveform on the fifth row of FIG.10.
On the other hand, when a different signature code is multiplied or when multiplication timing is not synchronized, the despread signal waveform becomes as shown by the signal waveform on the sixth row of FIG. 10. When an in-phase addition is applied to this signal, the correlation value becomes almost zero as shown by the signal waveform on the seventh of FIG.10 and no correlation is detected.
This is the explanation of in-phase addition (cumulative addition).
The procedure of the correlation detection calculation method of the present invention explained using FIG. 1 is summarized as shown in FIG.2.
That is, despreading is performed first before receiving a spread spectrum signal obtained by multiplying any one of a plurality of original signals (signature codes) with a specific cycle by a common spreading code (scrambling code) and detecting a correlation (step 20).
Then, a set of data bits corresponding to the cycle of the signature code which are in an equivalent positional relationship are extracted and a cumulative addition (in-phase addition) is performed on each set (step 21).
Then, each of the plurality of original signals (signature codes) is multiplied (step 22). Then, a cumulative addition is performed on symbol data (in-phase addition) for every multiplication and a correlation is detected (step 23).
When such a processing method is adopted, despreading is carried out by a common circuit, which suppresses the increase of the circuit scale.
Furthermore, carrying out a first cumulative addition (in-phase addition) focused on the periodicity of the signature code (original signal) which servers as the source of creation of a preamble signal reduces the number of symbols. More specifically, the number of symbols is' equal to the number of symbols making up the basic pattern.
Thus, a correlation is detected by carrying out multiplications of signature codes (parallel, time-division, simultaneous use of parallel/time-division, etc.) and finally carrying out a second cumulative addition (in-phase addition) while making the most of the hardware performance.
Thus, it is possible to attain the most efficient processing using minimum hardware. Furthermore, adjusting the configuration of multipliers or using a minimum memory on a time-division basis, etc. can further reduce the circuit scale and secure the flexibility of the circuit.
More specific advantages of the present invention will be explained in the following embodiments.

(Embodiment 2)

FIG.3 is a block diagram Showing a specific configuration example of the matched filter of the present invention.
For convenience of explanations, FIG.3 simplifies registers, etc. assuming that a signature code consists of repetitions of a basic pattern made up of two symbols of {an, bn}.
That is, the transmitting side (mobile station) creates preamble codes by multiplying signature codes consisting of repetitions of a basic pattern made up of two symbols of {an, bn} by common spreading codes (K1, K2, K3, K4···) and sends those preamble codes to the base station.
That is, in the following explanations, the signal obtained by spreading spectrum of the signature code having a two-symbol cycle at spreading factor "1" becomes the information for establishment of synchronization. Since the spreading factor is "1", one symbol has the same period as one chip. Then, the case where the base station carries out an in-phase addition over an 8-symbol period after despreading will be explained.
Furthermore, for convenience, suppose two kinds (1, -1) and (-1, 1) as the signature code . The first is expressed as signature code (1) and the latter is expressed as signature code (2).
Suppose such a, premise will also be applied in the following embodiments in the like manner. By the way, the above-described premise is simply intended for simplification for convenience of description, and the signature code actually has a length of, 16 symbols and there are 16 types of signature code.
Moreover, ,the value of a transmission symbol is normally expressed in, a binary number "0" or "1" and in response to this, suppose digital data takes a value "1" or "-1".
In FIG. 3, despreading codes (K1 to K4) generated by despreading code generation section 201 are stored in register 202. On the other hand, reception data (A to D) is temporality stored in register 204.
Then, multiplier 205a to 205d multiply the reception data by despreading codes first. Then, focused on the periodicity of the signature code, the symbol data at the same position of the basic pattern is extracted over two cycles and adders 206a and 206b perform cumulative additions (first in-phase addition).
This reduces the number of symbols down to the number equivalent to the number of times the basic pattern is repeated in the signature code. That is, it is possible to reduce the number of symbols without impairing the periodicity of the data string.
As a matter of course, additions are carried out, for example, every 16 symbols when the symbols data before spreading has a cycle of 16 symbols.
In the matched filter in FIG. 3, the above-described configuration constitutes the first correlation detection section for detecting a correlation about a common spreading code.
Output s:ignals of adder 206a and 206b are branched to two lines (originally 16 lines) to simultaneously multiply two kinds of signature code (signature codes (1) and (2)) in parallel.
Multiplication sections 220a and 220b are provided to multiply signature code (1) and signature code (2), respectively. One multiplier is constructed of code inverter 207a (207b to 207d) and selector 208a (208b to 208d) with two inputs and control section 209a (209b).
Control section 209a (209b) performs changeover control to select which of an inversion signal and non-inversion signal input to selector 208a 208b to 208d) according to the array of "1" and "-1" of the signature code to be multiplied.
That is, when the signature code is "+1", the non-inversion signal is selected and when the signature code is "-1", the inversion signal is selected.
The configuration of switching the selector according to the pattern of the signature code has a high degree of flexibility and is convenient because it is programmable according to the code to be multiplied.
Then, adder 211a (211b) performs a cumulative ad,d,ition on symbol data after being multiplied by the signature code (second in-phase addition).
The result of this second in-phase addition is stored in memory 212a (212b). This embodiment minimizes, the number of shift register stages and the, number of adders carrying out first and second cumulative additions to reduce the circuit scale, and, therefore cannot perform cumulative additions at a stroke by only the required number of symbols.
Thus, this embodiment requires time-division processing. For this reason; this embodiment provides small-volume memory 212a (212b) and allows this memory to store data little by little , circulates the, data and carries out second cumulative addition processing.
Input (write) and, output (read) timings of memories 212a and 212b are controlled by control signal CT from control sections 209a and 209b, respectively.
The section of the matched filter in FIG. 3 that multiplies the signature code and detects a correlation regarding the signature code constitutes the second correlation detection section.
FIG.4 shows the main processing content of the matched filter in FIG.3.
Since the number of shift registers in FIG.3 is 4, the number of symbols that can be added up at a time is 4.
Thus, to carry out an in-phase addition for 8 symbols, the calculation must be performed in two steps. In FIG.4, the first calculation is described as calculation section A and the second calculation is described as calculation section B. A case where signature code (1) is mutiplied will be explained below.
State ① in FIG. 4 shows a basic pattern (8 symbols) of signature code (1). A preamble signal is generated by multiplying these 8 symbols by spreading codes {1 ,1 ,-1,-1,1, -1,-1 ,1} (state ② in FIG.4: transmission signal).
This transmission signal (state ② in FIG.4) is input to the matched filter as is.
The matched filter multiplies the first half 4 symbols of the received code by spreading codes {1,1,-1,-1} to calculate calculation section A first. As a result, symbol data in state ③ in FIG.4, that is, data of {1,-1, 1,-1} is obtained.
This data is added up every cycle of the signature code, that is, every two symbols.
That is, in state ③ in FIG. 4, cumulative additions a+c and b+d are performed. The results are "+2" and "-2" (state ④ in FIG.4). These are equivalent to the outputs of adder 206a and 206b in FIG.3.
These addition results are multiplied by signature code (1). As a result, the outputs of selector 208a and 208b in FIG.3 are "+2" and "2" respectively (state ⑤ in FIG.4).
These outputs tare added up by adder 211a in FIG. 3 to become a value "+4" and stored in memory 212a.
Then a calculation of calculation section B is carried out using the same procedure. That is the last half 4 symbols (1,1,-1,-1) of the reception data are multiplied by despreading codes and the same processing as for calculation section A is carried out thereafter.
In the case of a calculation of section B, too, the output of adder 211a in FIG.3 becomes a value "+4" in the same way as the case of calculation of section A.
Control section 209a in FIG.3 reads "+4" stored in memory 212a as the result of calculation section A, adds it to "+4" the output of adder 211a of calculation section B to obtain "+8" and stores it in memory 212a again (state ⑥ in FIG.4). The matched filter of this embodiment can drastically reduce the circuit scaled .
FIG. 9 shows a configuration of a comparison example (conventional example). This comparison example (conventional example) adopts a configuration of providing correlation detection circuits (A1 to A16) in parallel by the number of types of signature code.
That is, codes are created by multiplying various signature codes by despreading codes and these are multiplied on the reception data simultaneously (in parallel) to determine a correlation value.
This configuration needs to provide correlation detection circuits corresponding to the number of types of signature code, thus increasing the circuit scale.
On the other hand, the configuration of this embodiment allows the section that carries out a multiplication of despreading codes to be used commonly, uses a hardware configuration of the minimum possible size, divides processing into steps in the time dircetion and carries out distributed processing as a flow production. Thus, this embodiment can fully reduce the hardware volume and attain low power consumption of the circuit.

(Embodiment 3)

FIG.5 is a block diagram showing a configuration of a matched filter according to Embodiment 3 of the present invention.
As in the case of the circuit explained in the aforementioned embodiment, the matched filter in FIG.5 is the same as the aforementioned embodiment in that it carries out a first in-phase addition according; to the cycle of signature codes after despreading the received data and then multiplies various types of signature code an carries out a second in-phase addition, to detect a correlation.
However, the circuit of this embodiment (case of FIG.5) despreads reception data, carries out a first in-phase addition (cumulative addition), temporarily stores in memory, and after calculations for a predetermined number of symbols are completed, carries out multiplications of signature codes and a second in-phase addition.
That is, this embodiment stores the data after the first in-phase addition in memory, adjusts the timing of writing to or reading from this memory and thereby absorbs a mismatch in the processing speed between the processing up to the first in-phase addition, multiplications of,signature codes and the processing of the second in-phase addition and is different in this respect from the circuit in FIG.3.
This configuration can implement a desired correlation detection calculation with a smaller circuit scale than the circuit (FIG.3) in Embodiment 2.
For convenience of explanations, as in the case of the above-described Embodiment, the following explanation will describe a case where the spreading factor is 1x, a signal on which an original signal (signature code) with a two-symbol cycle is spread is despread and then subjected to an in-phase addition over a period of 8 symbols.
Configurations and operations of reception data storage shift register 204, despreading code generation section 201 and despreading code storage shift register 202 are the same as those of the circuit in FIG.3.
The results of additions of adder 206a and 206b are directly input to memories 301a and 301b and temporarily stored unlike Embodiment 2.
In order to perform an in-phase addition for an 8-symbol period, calculations for 4 symbols are repeated 2 times as in the case of Embodiment 1. Since reception signal storage shift register 204 and spreading code storage shift register 202 have a 4-stage configuration, despreading codes are newly set twice.
With the second despreading, the spreading codes' are fixed and input data is shifted likewise Then, the selectors (not shown) are switched for every cycle of a signature code and a first in-phase addition (accumulation) is performed and the result is stored in memories 301a and 301b.
With the second calculation for every 4 symbols, the moment data is output from adders 206a and 206b, data of the first calculation result is read from memories 301a and 301b and the data is added up by adders 206a and 206b and stored in memories 301a and 301b again.
After a calculation for a predetermined number of symbols is completed, data is read from memories 301a and 301b, multiplied by signature codes and a cumulative addition is performed on symbol data (second in-phase addition).
Control section 302 reads data from memories 301a and 301b. Then, while one portion of data is being read, control section 302 multiplies signature code (1) and signature code (2) consecutively.
In the case where the number of types of signature code is 16, these 16 types of signature code (signature code (1) to signature code (16)) are, for example, multiplied consecutively at a stroke.
Multiplications of signature codes are performed by selectors 208a and 208b changing between inversion and non-inversion data according to the bit array of signature codes. This embodiment is the same as Embodiment 2 in this respect.
Then, adder 211a performs a cumulative addition on symbol data portions resulting from the multiplications of signature codes (second in-phase addition).
Using a method of switching between the selectors for' two times of signature codes (1) and (2) while one portion of data is being read from memory allows the correlation detection results of signature codes (1) and (2) to be output alternately.
When thee lock frequency of reading data from memories 301a and 301b is assumed to be "A"', the clock frequency of switching between selectors 208a and 208b is "2A" (first reading method).
The above-described explanation describes the case where a correlation of all signature codes is detected while data is being read from memory once, but the present invention is not limited to this.
There can also be a method whereby data is read from memories 301a and 301b twice and a multiplication and' in-phase addition of signature code (1) are carried out in the first read and a multiplication and in-phase addition of signature code (2) are carried out in the second read (second reading method).
In the case of the first reading method, when the number of types of signature code increases, the selector needs to be changed at a faster clock.
On the other hand, in the case of the second reading method, the same clock can be used for, reading from memory and selector- changeover, but the time required to obtain the correlation result is longer than the first, reading method. The selection of a reading method must be determined according to the function required.
This embodiment selects the number of times memory is read and the selector changeover speed according to limits of the hardware processing speed as appropriate and flexibly carries out time-division, processing, and can thereby reduce the circuit scale.

(Embodiment 4)

FIG.6 is a block diagram showing a configuration of the matched filter of the present invention. A circuit configuration and operation are basically the same as those in Embodiment 3 (matched filter in FIG.5).
That is, this embodiment adopts the system of storing data in memory after a first in-phase addition and then carries out multiplications of signature codes and a second in-phase addition.
In the case of the matched filter in FIG. 5, which performs serial processing, the hardware processing cannot catch up with an increase in the number of types of signature code (code combination) causing an increase in the number of times of time-division, which tends to slow down the processing speed.
Therefore, this embodiment adopts concurrent use of parallel processing and time-division processing (combined parallel/time-division system) in order to maximize the hardware processing capacity and attain efficient processing.
The number of parallel processes is determined based on the type of signature code (number of combinations) and hardware processing speed. Assuming that a maximum of "m" signature codes can be multiplied while data is being read from memory one time, if the number of types of signature code (number of (combinations) is "n", then the number of parallel processes p is determined by "n/m".
As illustrated in FIG. 6, if the number of types of signature code is 16 (signature codes (1) to (16)) and the maximum changeover frequency of selectors 208a to 208d is 4 times the read clock frequency of memories 301a and 301b, the number of parallel processes is "4 (=16/4)".
That is, parallel processes of multiplying 4 different signature codes are carried out simultaneously.
Therefore, according to the circuit in FIG.6, by one read from memories 301a and 301b, it is possible to perform correlation detection processing' of a plurality of signature codes (multiplication processing and in-phase addition processing) in parallel, thus improving the processing efficiency.
The processing of the method in Embodiment 3 alone may be limited, and as in the case of this embodiment, it is effective to combine parallel processes according to the processing speed and processing time required for processing as appropriate.

(Embodiment 5)

FIG. 7 is a block diagram showing a configuration of a CDMA communication base station apparatus using the matched filter bf the present invention.
An overall configuration and operation of FIG. 7 are the same as those explained initially in Embodiment 1.
The spread spectrum communication system normally sends signals with phase modulation applied for a radio section. For this reason, demodulation circuit 401 on the receiving side demodulates the received signal subjected to phase modulation first. The demodulated signal is converted to a digital signal through A/D converter 402.
The converted digital signal is input to matched filter 403a. As explained in Embodiments 1 to 4 above, the matched filter is constructed of reception data storage section 500, despreading code storage section 502, despreading code generation section 501, memory 505 and multiplication operation section (including in-phase addition circuit), etc and carries out despreading and in-phase addition on a plurality of signature codes. Correlation detection of each signature code is performed in this way.
Based on the calculated correlation value, power is calculated and a delay profile is created. The matched filter normally carries out despreading and in-phase addition corresponding to the number of symlibls requested by the system up to the delayed reception signal.
By creating a delay profile, it is possible to determine which signature code is received with what amount of delay. Thus, timing information necessary for establishment of synchronization is obtained for every signature code (preamble search).
The synchronization information acquired at this time is used to establish synchronization for giving communication permission to the transmitting side based on the spread spectrum communication system.
That is, when a search with a preamble signal (preamble search) is completed, this is followed by 'a' demodulation search for sampling a reception signal with a narrower time width than this preamble search.
The above-described operation of the base station can be summarized as shown in FIG.8.
That is, despreading is performed by multiplying the received I signal and Q signal by spreading codes (step 700).
Then, focused on the periodic array of signature codes (original signals with a certain period), a set of symbols at the corresponding positions (at the same position in one cycle) are extracted and subjected to a cumulative addition (step 701).
Then, signature codes are multiplied and the resultant symbol data is subjected to a cumulative addition and then a second in phase addition (step 702).
Then, I²+Q² is calculated (power calculation), compared with a threshold and judged whether a preamble signal has been received or not and a delay profile is created for every signature code.
Then, the reception signal is sampled with a narrower time width than a preamble search to carry out a demodulation search) (step 703).
Using the present invention can drastically reduce the scale of the circuit that carries out a preamble search at the base station apparatus. This can also reduce power consumption of the circuit.
FIG. 9 shows an example (conventional example) of a circuit that carries out correlation calculations in parallel for every signature code.
As described above, this circuit generates codes obtained by multiplying each signature code by a common despreading code in parallel and multiplies the reception data by these codes simultaneously. Such a circuit configuration requires the same number of despreading circuits as the number of types of signature codes, causing the circuit scale to become enormous. Moreover, since the hardware configuration is fixed, this circuit configuration' is liable to incur waste.
On the other hand, the matched filter of the present invention performs despreading using a common circuit and can thereby suppress an increase of the circuit scale. The matched filter of the present invention reduces the number of symbols by periodically adding portions of data with a same value (first in-phase addition) focused on the periodicity of signature codes (original signals) which serve as the source of creation of a preamble signal, carries out multiplications of signature codes while maximizing the hardware performance (parallel, time-division, combined parallel/ time-division) and finally carries out a second in-phase addition to detect a correlation.
Thus, the matched filter of the present invention can provide the most efficient processing making full use of minimum hardware. It can also adjust a multiplier configuration, use a minimum memory on a time-division basis, etc. to further reduce the circuit scale thus securing the flexibility of the circuit.
The above-described explanations describe preamble, search processing, but the present invention is applicable to processing for establishment of initial synchronization using pilot signals which are known non-modulated signals (can be regarded as periodic signals) as well.
As explained above, when a correlation needs to be detected instantaneously for each of a plurality of periodic signals included' in' a reception signal as in the case of a preamble search at a CDMA communication base station, the present invention can provide processing with maximum efficiency with a small circuit scale. Thus, the present invention, realizes low power consumption of the circuit as well.
The present invention is not limited to the above described embodiments, and various variations and modifications may be possible without departing from the scope of the present invention.
This application is based on the Japanese Patent Application No.2000-341475 filed on November 9, 2000

Claims

A matched filter in a base station within a CDMA communication system for detecting a preamble, which is obtained by multiplying a signature code selected by a mobile station from a plurality of available signature codes consisting of n repetitions of a basic pattern by a common spreading code, sent from the mobile station (1) to the base station (2), comprising:
a despreading means (501, 502, 503) for despreading the received preamble sequence by multiplying it with the common spreading code;

characterized by

a first cumulative addition means (3) for performing the cumulative addition on the despread signal on a per symbol basis, wherein a set of n symbols ("a_n", "b_n", "c_n", "d_n") at the same position one the basic pattern of the despread signal is extracted over the n repetitions and subjected to the first cumulative addition;

a multiplication means (4) for multiplying the output of the first cumulative addition means (3) with each of the plurality of available signature codes on a per symbol basis respectively; and

a second cumulative addition means (5) for performing a cumulative addition on the output of the multiplication means (4)
The matched filter according to claim 1, wherein the multiplication means (220a, 220b) is further configured to comprise:
a plurality of code inverters (207a, 207b) for inverting the output of the first cumulative addition;

a plurality of selectors (208a, 208b) for selecting a non-inverted signal or an inverted signal according to the control of a control means (209a); and

a control means (209a) for performing the changeover control to the selecting means (208a, 208b) according to the polarity of each symbol of the available signature code.
The matched filter according to the claim 1, further comprising:
a plurality of storing means (301 a, 301 b) for temporarily storing the output of the first cumulative addition means;

a plurality of code inverters (207a, 207b) for inverting the output of the first cumulative addition;

a plurality of selectors (208a, 208b) for selecting a non-inverted signal or an inverted signal according to the control of a control means (209a); and

a control means (302) for multiplying each available signature code with the output of the first cumulative addition means respectively on a per symbol basis, while reading one portion of data from the storing means (301a, 301 b).
A preamble detection method used in a CDMA communication system, for detecting a preamble, which is obtained by multiplying a signature code selected by a mobile station from a plurality of available signature codes consisting of n repetitions of a basic pattern by a common spreading code, sent from the mobile station (1) to the base station (2), comprising the steps of:
despreading the received preamble sequence by multiplying it with the common spreading code;

characterized by

performing the cumulative addition on the despread signal on a per symbol basis, wherein a set of n symbols ("a_n", "b_n", "c_n", "d_n")

at the same position of the despread signal is extracted over the n repetitions and subjected to the first cumulative addition;

multiplying the output of the first cumulative addition with each of the plurality of available signature codes on a per symbol basis respectively; and

performing a cumulative addition on the output of the multiplication.
A base station including a matched filter as defined in any of claims 1 to 3.